RAG (Retrieval-Augmented Generation) for GIS + Documents

VenkataKondepati — Fri, 05 Sep 2025 03:50:40 GMT

Problem: Many organizations want to ask natural language questions that combine GIS layers with documents (plans, zoning PDFs, reports). Today, GIS search is either keyword- or map-based, not semantic.

Enhancement: Add a “Semantic Search Assistant” in ArcGIS Hub and Enterprise that uses a vector database behind the scenes to search across:

GIS layers (attributes + geometry summaries)
Documents (PDFs, DOCX, etc.)
It would return both map features and document excerpts, with citations.

Use Case: A city planner could ask: “Which neighborhoods are zoned for mixed use, and what are the design guidelines?” → System returns zoning polygons on the map + excerpts from the design guideline PDF.

idea RAG (Retrieval-Augmented Generation) for GIS + Documents in ArcGIS Survey123 Ideas

RAG (Retrieval-Augmented Generation) for GIS + Documents